Copy Number Variations

Transcription

1 Copy Number Variations Illumina Seminar - Milan June 18, 2009 Untangling the complexity of mendelian and complex diseases Federica Torri Dept of Science & Biomedical Technologies Fondazione Filarete, University of Milan

2 What are CNVs? Stretches of DNA larger than 1 kb that display copy number differences in comparison to a reference genome - Copy number variation (germline,, inherited) - inherited: also present in parents genome - de novo: absent in parents genome - Copy number alteration (somatic, e.g. in cancer cells) - Copy number polymorphism (relatively common CNV, with a fixed starting/ending position) - Copy number difference (between-species copy number differences, e.g chimpanzees and humans)

3 Nat Rev Genet. 2007;8: CNVs in α-globins CNVs in CMT CNVs as a general feature

4 Structural variation in the human genome Nat Rev Genet 7: 85-97, 2006 Millions of nucleotides of heterogeneity within every genome likely to make an important contribution to human diversity and disease susceptibility. ~12 % or 4.8%?

5 CNV: the hottest topic?

6 The current map of human structural variation is far from complete.

7 CNVs in health Important in primate evolution Overlap and regulate the expression of ~ 18% of genes Account for more nucleotide variation than SNPs in the human genome Non-random distribution across the genome: hot spots Nearly 50% map in SD s regions Mechanisms: NAHR, NHEJ, retrotransposition, FoSTeS Underwent selective pressure in the human lineage (shaped by mutation, selection, demographic history) Population-specific variation Enriched towards environmental sensor genes (sensory perception and defense response functions) AMY1 gene paradigm

8 CNVs in health Important in primate evolution Overlap and regulate the expression of ~ 18% of genes Account for more nucleotide variation than SNPs in the human genome Non-random distribution across the genome: hot spots Nearly 50% map in SD s regions Mechanisms: NAHR, NHEJ, retrotransposition, FoSTeS Underwent selective pressure in the human lineage (shaped by mutation, selection, demographic history) Population-specific variation Enriched towards environmental sensor genes (sensory perception and defense response functions) AMY1 gene paradigm

9 Population differentiation Redon et al., Nature ;444:444-54

10 CNVs in health Important in primate evolution Overlap and regulate the expression of ~ 18% of genes Account for more nucleotide variation than SNPs in the human genome Non-random distribution across the genome: hot spots Nearly 50% map in SD s regions Mechanisms: NAHR, NHEJ, retrotransposition, FoSTeS Underwent selective pressure in the human lineage (shaped by mutation, selection, demographic history) Population-specific variation Enriched towards environmental sensor genes (sensory perception and defense response functions) AMY1 gene paradigm

11 CNVs and evolution: how has CNVs shaped our genomes? Genome Res. 2008, 18: CNVs number and median size comparable orthologous region significantly enriched human specific SDs significantly enriched for human-specific CNVs, the same for chimpanzee: species specific CNVs pattern due to architectural differences between the two genomes Comparison between fixed CNDs (human and chimp) and human CNVs containing genes.

12 The relations between CNVs affecting genes and fixation future studies aiming to characterize genetic basis of adaptive phenotypic differences between humans and chimpanzees.

13 latest news on CNVs differences between any two individuals 99% of the copy number differences between any two individuals is explained by the simple mendelian inheritance of the same allele from a parent COMMON VARIANTS -Only few >100kb -Copy-number differences between two individuals seems to be due to CNPs: 90% from CNPs with MAF > 1%, 80% from CNPs with MAF > 5% -mostly inherited -patterns of allele frequency, LD, population differentiation that mirror SNPs -90% biallelic -10% multiallelic RARE VARIANTS -Mostly >100kb -Affect less than the 5% of the human genome - Rare CNVs sharing imply sharing of larger genomic region -fewer allelic states -highly penetrant and short lived in the population -De novo/inherited persisting for only a few generations within a pedigree next frontier of human genetics research!

14 1000 genomes projects: a glimpse on lower frequency variants -AIM : a nearly complete catalog of common human genetic variants (defined as frequency 1% or higher) by high-quality sequence data for >85% of the genome for three sets of individuals - Each set is sampled from broad geographic regions (tentatively, Bantu-speaking populations in Africa; populations in East Asia; and populations in Europe). -This catalog will include SNPs, copy number variants, and short insertion and deletion polymorphisms. Filling the gap between rare and common variants

15 Am J Hum Genet. 2009;84: ~16% of the genome Low frequency High frequency

16 CNVs not only in simple (genomic) diseases Nat Genet 40: 322-8, 2008 GENOMIC DISORDERS: sporadic diseases caused by recurrent de novo structural mutations

17 but also in complex traits McCarroll SA,Hum Mol Genet ;17:R135-42

18 The common versus rare hypothesis Common variant common disease scenario Rare variant common disease scenario vs Case-control CNVs mapping in shared regions AIM : to find CNVs segments that are associated to a complex trait in our cohort, looking to replicate this findings in other samples. SUGGESTED ASSOCIATION ANALYSIS: difference in allele frequency between cases and controls Rare CNVs mapping in multiple regions defining an individual pattern of variation AIM : to find CNVs patterns that may be differentially associated to a complex traits (looking for a shared biological function) SUGGESTED ASSOCIATION ANALYSIS: enrichment of a collection of rare variants in cases or controls

19 The schizophrenia paradigm Mixture of common and rare variants likely Mendelising pedigrees uncommon - no common risk variants found to date - mathematical modelling of familial risk is inconsistent with single genes of large effects Apparently stable incidence and prevalence Reduced fecundity Multiple rare recurrent CNVs that underlie strong negative selection may account for a larger fraction of overall genetic risk Common variants Rare variants small genetic effects POWER PROBLEM rare sequencing needed at the single nucleotide/small structural level

20 walking through complexity Owen MJ et al., Curr Opin Genet Dev. 2009

21 ..converging evidences on loci Chr. 1q21.1 Chr. 15q11.2 Chr. 15q13.3

22 ..converging evidences on genes

23 genomic burden of rare variants Science 320: , 2008 Nat Rev Genet 40: , 2008

24 Detecting CNVs is not a trivial issue - Circular Binary Segmentation (Olshen et al., 2004) - cnvpartition (Illumina Inc., San Diego, CA) - Circular Binary Segmentation Modified -Nexus Copy Number (Biodiscovery Inc., El Segundo, CA) - Hidden Markov Models (HMMs) - QuantiSNP (Wellcome Trust for Human Genetics, University of Oxford) - PennCNV (University of Pennsylvania) - Golden Helix Dynamic Optimal Segmentation (Golden Helix Inc.)

25 Nexus v4 (Biodiscovery Inc.) All Chromosome overview CNV detection Whole genome overview All Chromosome visualization Single Chromosome visualization Numerical data on the aberrations

26 Threshold for Nexus parameters Log R ratio homozygous gain heterozygous gain heterozygous loss Copy number evaluation homozygous loss BB AB AA B allele frequency Allelic imbalance LOH Allelic imbalance Allelic imbalance Allelic imbalance evaluation Normal zygosity

27 B-allele frequency possible scenarios Father: AABABBBABABBBAAABABB Mother: ABBABA ABABBAAA A AA AB BB AA BB AB B A B A B B AB AB AAA AAB BBB AAA ABB ABB BB AB AA Normal LOH Imbalance 100% 0% BBB BBA BAA AAA

28 Our approaches to CNVs topic 1) Schizophrenia : exploring the two different sides 2) Validation of a de novo genomic loss of heterozygosity (LOH) for HLA haplotype as a mechanism of in vivo leukemia immune escape in patients with relapse after haploidentical transplantation and adoptive transfer of donor T cells 3) Novel gene discovery in monogenic cancer predisposition syndromes: a first proof of concept of our analysis pipeline

29 Our approaches to CNVs topic 1) Schizophrenia : exploring the two different sides 2) Validation of a de novo genomic loss of heterozygosity (LOH) for HLA haplotype as a mechanism of in vivo leukemia immune escape in patients with relapse after haploidentical transplantation and adoptive transfer of donor T cells 3) Novel gene discovery in monogenic cancer predisposition syndromes: a first proof of concept of our analysis pipeline

30 Schizophrenia : exploring the two different sides Common variant common disease scenario Rare variant common disease scenario vs n= 89 n= 117

31 Nexus v4 (Biodiscovery Inc.): the comparison feature CNV detection Frequency Comparison COMPARISON FEATURE: comparison in CNVs frequency between groups assessing which CNVs are significantly enriched

32 are we moving in the right direction? Mol Psychiatry Only SNP replicated maps in DOCK4 DOCK4 activates Rac GTPase Predominant expression in hyppocampus Upregulated at the same time dendrites start growing Knockdown results in impaired dendritic morphogenesis

33 Our approaches to CNVs topic 1) Schizophrenia : exploring the two different sides 2) Validation of a de novo genomic loss of heterozygosity (LOH) for HLA haplotype as a mechanism of in vivo leukemia immune escape in patients with relapse after haploidentical transplanation and adoptive transfer of donor T cells 3) Novel gene discovery in monogenic cancer predisposition syndromes: a first proof of concept of our analysis pipeline

34 Genomic loss of mismatched HLA in leukemia is a major mechanism of in vivo escape from T cell immunosurveillance following haploidentical HSCT Vago L. et al., Genomic Loss Of Mismatched HLA In Leukemia is a Major Mechanism of in Vivo Escape from T Cell Immunosurveillance Following Haploidentical HSCT, NEJM, in press

35 Our approaches to CNVs topic 1) Schizophrenia : exploring the two different sides 2) Validation of a de novo genomic loss of heterozygosity (LOH) for HLA haplotype as a mechanism of in vivo leukemia immune escapein patients with relapse after haploidentical transplanation and adoptive transfer of donor T cells 3) Novel gene discovery in monogenic cancer predisposition syndromes: a first proof of concept of our analysis pipeline

36 Gene discovery: looking for a needle in a haystack knowing the needle to search for Reporter pathogenic CNV in a monogenic cancer predisposition syndrome (HNPCC) = MLPA % of patients are negative for mutation/rearrangements in candidate genes and display a phenotype superimposable to known monogenic cancer predisposition syndromes NFMP-HPLC + Human1m-duo BeadChips pathogenic CNV boundaries ortogonal validation empirical testing of theoretical parameters to optimize analysis conditions

37 Gene discovery : towards the gene point of view CNV detection CNVs list Frequency Comparison CNVs regions present only in cases 120 genes Gene filtering needed!!!!

38 Gene discovery : towards the gene point of view Gene filtering Cumulative Gene Region IDs Region Code Explicited Cumulative Freq Cases Freq Controls GKAP1 0 chr9:85,616,440-85,623,279 7,5 0 CKMT1B 5,031,166 chr15:41,676,268-41,682,935 <-> chr15:41,660,781-41,676, ANO3 0 chr11:26,565,073-26,567, MSH2 702,0 chr2:47,460,187-47,511,188 <-> chr2:47,539,880-47,547, LOC chr1:147,779, ,849, PPIAL4C 0 chr1:147,779, ,849,

39 Gene discovery : identification of novel cancer predisposition genes???????? Mutation negative in candidate genes patients cohort (20-55 %) CNV detection? Frequency Comparison Gene filtering

40 CNVs genomic content is different between Monozygotic twins: frequency of differences in genetic makeup up to 10 %!!! Am J Hum Genet. 2008;82: Different tissues within an individual (mosaicism) Notini AJ et al., Cyt Genome Res : Piotrowski A. et al.,. Hum Mutat. 2008;29:

41 Many-but far from all-common CNVs can be interrogated with a tag SNP on genotyping platforms Nat Genet 40: , % of the common deletions (worldwide frequency >5%) strongly correlated to a HapMap SNP (worldwide r 2 > 0.8) Nat Genet 40: , 2008 More probes but more background rumor High density genome-wide SNP array tagged about ~50% of the common deletions Segmental duplication under-representation or cross-hybridization of paralogous sequences NAHR-driven mutation less likely to be tagged LD with surrounding SNPs: underestimation

42 Towards deep sequencing. Nature 453: 56-64, % in more than 5 ids 1695 sites of structural variation: 49 % more than 1 ids, 15% more than regions of novel insertion sequences not present in the reference genome: 40% are CNVs ranging (from few kbp up to 130 kb) nearly 50% lay outside previously reported CNVs region Comparison with the highest density SNP commercial platform (Illumina 1M and Affy6): 50% deletions cannot be detected! DE NOVO SEQUENCING OF ADDITIONAL GENOMES NEEDED!

43 Looking to individual structural variants with sequencing technologies PLoS Biol ;5:e million DNA variants (~12.3 Mb) (~30%) novel! 3,213,401 single nucleotide polymorphisms (SNPs( SNPs) 53,823 block substitutions ( bp) 559,473 homozygous indels (1 82,711 bp) 62 CNVs 292,102 heterozygous insertion/deletion events (indels( indels)(1 571 bp) 90 inversions Non-SNP DNA variation accounts for 22% of all events identified in the t donor, however they involve 74% of all variant bases

44 Assessing the relevance of CNVs in clinic Cook et al., Nature ;455:919-23

45 Aknowledgements Laboratory of genetic of complex traits, University of Milan Erika Salvi Sara Lupoli Cristina Barlassina Luisa Strik Lievers Valentina Tieran Francesca Taddeo Cristian Cosentino Fausta Schiavini Daniele Cusi Fabio Macciardi Department of Oncology and Neurosciences, University G. D Annunzio, Chieti Laura De Lellis Alessandro Cama Istituto Nazionale dei Tumori, Milan Paolo Radice IRCCS San Raffaele Luca Vago Fabio Ciceri Katharina Fleischauer Istituto Tecnologie Biomediche, CNR Alessandro Orro Andrea Calabria Paolo Cozzi UC Irvine, CA Steven G. Potkin Jessica Turner James Fallon Guia Guffanti Illumina, Inc. Mita Mancini Chiara Dal Fiume Biodiscovery, Inc. Soheil Shams